Bayesian boiler challenge
Posted by:
ianc (IP Logged)
Date: April 23, 2010 01:28AM
I don't know if anyone reading this is familiar with Bayesian statistics, frequently called 'decision making under conditions of uncertainty', but it was developed long before Karl Pearson made himself famous by saying, essentially, "why bother with uncertainty? We can measure variables, so we can predict what the result will be." Unfortunately, the soft sciences as well as the hard heeded his siren call, which is why we can put a man on the moon but we can't predict what the stock market will do. What has this to do with boilers? Well, after getting my boiler to stop weeping at street pressure (110 psi) I rebuilt my water hand pump, connected it, and put 200 psi on the boiler. It doesn't weep. (Course, now that I've said that....) So I've looked at the advice I've had on how to proceed, and here's what I've got.
(1) Once you start the hydo testing, go in 100# intervals. I always go to 700# of water pressure. Keep a towel to dry the boiler and check for drips. If the bottom is wet but no drips, don't worry about them, they will seal up. If there is a major drip, re swage it only very very very lightly, about 1/2 as much force as you swaggd the rest of the tubes, you only want to expand the tube just a little bit. Remember to take it up to steam slowly.
(2) Until you can pressurize that boiler at 25% above operating pressure and hold it with a hydrostatic test, I wouldn't be using that boiler in service.
(3) Take it up to 300 # and see what happens. If they are not dripping, they will probably seal up once there is fire. It is better to not swage too much and have to go back later and swage the boiler again than to swage too much and have to replace a boiler.
(4) If they are not leaking then hydro test the boiler using the hand water pump in the car. I hydro test my boiler tubes at 400 psi cold. They should all be pretty tight top and bottom at 400 psi cold. If not, swage them again, row by row until they all are tight. both top and bottom. In the end, if you do have a couple weeping tubes. they can be addressed individually with the swaging tool. At 400 psi cold, a drip every 15 seconds from a tube would be barely acceptable. On a hot boiler, they would not leak at all because of the expansion rate of copper being so much greater than the steel tube sheet.
(5) If it was me and those are only weeps at 100 psi, I would be inclined to steam up as weeps will plug themselves in most cases. There is always a portion of debris or mud in the boiler and its all it takes to finish the job. On new retube, "bars leak" in the boiler and roll it around under pressure, will fix small weeps, I would even do that on a used boiler before giving up to retube, as it can always be flushed out if it seals up. Any small amount of sealer not flushed won't hurt anything, so long as the cylinder oiler is working to the engine.
So. Entering all that into Bayesian equation, I find the probability of making the correct decision is exactly 50%. That is, a coin toss. And by the way, if I should take it above 600 PSI, given that the pressure relief will go off at 650, do I adjust it so it won't and then reset it?
Thanks, from a puzzled statistician.